Image processing apparatuses such as an electronic camera which uses a memory card having a solid-state memory element as a recording medium, and records and plays back still and moving images sensed by a solid-state image sensing device (to be described as an image sensing device hereinafter) such as a CCD or CMOS have commercially been available. This image processing apparatus such as an electronic camera allows the photographer to select a single-shot/sequential image sensing mode from the operation unit. The photographer can switch image sensing between single-shot image sensing for sensing an image for each frame every time he/she presses the shutter button and sequential image sensing for sequentially sensing images while he/she keeps pressing the shutter button.
To sense an image by using the image sensing device such as a CCD or CMOS, the image processing apparatus can execute dark noise correction processing by calculation processing using dark image data read out after charge accumulation similar to actual image sensing while the image sensing device is not exposed, and image data of actual image sensing read out after charge accumulation while the image sensing device is exposed. A high-quality image can be attained by correcting the sensed image data for image quality degradation caused by dark current noise generated by the image sensing device, a defective pixel due to a slight scratch unique to the image sensing device, or the like. However, in order to cause the image processing apparatus to perform dark noise correction processing, a dark image must be sensed. This increases the release time lag, missing a good opportunity of capturing an image.
To solve this problem, there is known an image processing apparatus which uses correction data stored in advance to cancel the horizontal shading (luminance level nonuniformity) of the image sensing device or a noise component (offset from a proper dark level) such as a dark current, and can sense a high-quality image while suppressing the release time lag small.
The correction data stored in advance is an offset amount for canceling the horizontal shading of the image sensing device, or the difference between a proper dark level and image data obtained by performing dark image sensing but not performing correction using correction data in assembling an image processing apparatus. The dark level serves as a criterion for the luminance component and color components of image data in image processing. The image quality can therefore be improved by correcting the dark level of image data obtained by exposing the image sensing device.
The prior art suffers the following problems. Some image sensing devices nonlinearly change the dark current noise state depending on the temperature characteristic of an output circuit. In an image processing apparatus having such an image sensing system, a noise component which should be canceled remains in sensed image data even by using a correction value stored in advance, degrading the image quality. In this case, correction by calculation using a temperature coefficient complicates the calculation. Calculation processing takes a long time in the presence of many pixels, increasing the release time lag. A correction value may be stored in advance for each temperature region, which requires a larger memory capacity and makes the apparatus bulky.
In addition to dark noise correction processing, the image processing apparatus can execute shading correction processing by calculation processing using shading correction data stored in advance in a storage medium, and sensed image data read out after charge accumulation while the image sensing device is exposed. Noise generated in an image sensing circuit system, i.e., the voltage nonuniformity caused by the resistance component of the power line in a sensor, and shading by element variations or the like can be reduced, sensing a high-quality image.
However, the prior art poses the following problems. In a conventional image processing apparatus such as an electronic camera, shading correction data is stored in a storage medium in advance. In image sensing, the shading correction data is read out from the storage medium, and calculation processing is performed using the shading correction data and sensed image data, achieving shading correction. If the change of the shading amount depending on image sensing conditions is not considered, appropriate shading correction cannot be done, and the image quality may degrade. If the change of the shading amount depending on image sensing conditions is considered, the number of shading correction data corresponding to respective image sensing conditions must be stored in the storage medium, which requires a large-capacity storage medium.